aiLL 7/7/23 - 2 - The Most Important AI Tool in Human History?

Transcript

1. 0:15yeah all right I'm coming back yep is back welcome yep yep welcome XXX all right people are coming back if
2. 0:24you would go ahead thank you very much so CN is so was was that just some un trolling me that said that cuz I've been
3. 0:31booted here twice but it just booted me it just booted me and then said you can't come back for 7 days I never got
4. 0:38anything like that okay all right that's fine I figured that but like because I'm so far behind on my things like I didn't
5. 0:45know what it was oh you know those are trolls okay good great all right thank you people you know it's again this is
6. 0:52this is like you know remedial remedial old guy Tick Tock lessons anyway go ahead and tap the screen uh
7. 1:04let's get some likes going and maybe you can share the live so so people can do it thanks for the rose that's great um
8. 1:11you know those are trolls no I didn't know those were trolls because it it looked like it had a Tik Tock um icon
9. 1:19but I guess someone could just put a Tik Tock icon there yeah I get it whatever what are you going to do we
10. 1:27we'll we'll have a we'll have a more intimate grouping um cuz cuz what did we get up to we were up to like 130 people
11. 1:34there for a while so maybe we'll get some more people go ahead and tap the screen share the live let me know
12. 1:40encourage more viewers to share yeah viewers share would you would you share would you do that would you would you
13. 1:47help an old man all I've ever wanted was a Tik Tock following I came on the Tik Tok I I just wanted to be loved I just
14. 1:58wanted to be loved d Can I Be Loved won't you share this live oh yeah there's like 163 people here holy crap
15. 2:06okay when I'm talking like an idiot it's one of the things that makes me happy that people are like scrolling through
16. 2:12Tik Tok and there's just this weird ass old man going ah yes I was hoping that people would like me someday but anyway thank you um oh I'm
17. 2:25supposed to call it the Ticky talk yeah welcome to the Chickity talk Sally Sally how do I turn off the tick
18. 2:34talk I see my face again what do I do I I I don't want to see my face it scares have you seen my hair Sally Sally yeah
19. 2:44exactly all right love your energy thank you very much you'll be permanently banned that's a shitty troll all
20. 2:53right cuz I didn't do anything and if they wanted to if they were going to boot me they had just booted me all
21. 2:59right I'll get there I'll figure this out anyway tell me is prompt engineering an actual job it is and it's it is and
22. 3:06it's not and it's going to be and it's not so um it's a job for any company trying to figure out what the [ __ ] to do
23. 3:15with AI um they need people that understand how to talk to these tools and right now these tools are so the troll one at least temporarily I
24. 3:26know let's go back to the content what content was I doing tell me what content I was doing I'll go back to it cuz cuz you know I got thrown
25. 3:35I was like I was like you're going to pay me burm permanently I can't lose my Ticky talkie Ticky talky Elder talk this is it
26. 3:49welcome old guys trying to twerk that notification that'll get you banned is a troll account those son of a [ __ ] sons of
27. 4:02[ __ ] uh we can go back to prompt engineering is a job let's go back to the content I don't know what the
28. 4:07content was to tell me what the content was was there content I know that's where I am Amelia's wife I don't think
29. 4:14there was content you where you talk about the cost oh oh okay so I was talking about okay so so if you're new here my
30. 4:24name is Kyle Shannon this is the AI learning lab uh these URLs here are just here really for for your it's like a
31. 4:31public service announcement if you haven't played with chat GPT go to chat. open.com that's the official website of
32. 4:39chat GPT just go there and start playing and if you don't know what to do there go to prompt. chat and that's a document
33. 4:44that will tell you how to do it I'm not affiliated with them I don't make money I don't give a [ __ ] anyway okay so what
34. 4:50we were talking about before was you know someone asked about like you know how much is the you know what do I think
35. 4:56the server cost is or the the like the hosting cost um it's tremendously expensive to to do for for us to be able to play with these
36. 5:07like it is we are in a very very very very lucky position as non people working at these companies to have access to these tools
37. 5:21um the fact that chat GPT went from Z to 100 million users in 6 weeks starting last November changed everything
38. 5:33right Microsoft was not expecting it to be that popular open AI was not expecting it to be that popular Google
39. 5:41was certainly not expecting it to be that popular but but when it did what it did everyone started taking this
40. 5:49seriously Microsoft went from a billion dollars to1 billion doar and and basically basically said we're going to support this
41. 5:58thing and we're going to the the interesting thing about Microsoft as compared to Google so Google all they
42. 6:06have is their ad Revenue they've got you know they've got their search ad Revenue Microsoft has their entire software product Suite that
43. 6:16they're incorporating chat GPT into so they' got a completely different way to monetize it and the CEO of Microsoft has
44. 6:22said he's willing to lose money on chat GPT as part of Bing to [ __ ] with Google and he said it with some Glee he didn't
45. 6:31use those words but you know that's what he was saying um and also they're going to be able to monetize this through you
46. 6:37know an additional monthly charge for Office 365 and for Windows and for all the Azure cloud services so they're
47. 6:44they're going to make their money back on the business services and the Consulting and all that other [ __ ] that
48. 6:48they do so they can take a loss leader on this front but the fact that it got so popular is why everyone is is sort of
49. 6:57piling onto this AI stuff why it's got so much hype um it's also I think the reason that meta let their llama model
50. 7:07out into the wild um really interesting move there they they said it was a mistake you know it's a leak it was it
51. 7:14wasn't really a leak they're like well we give it to a few developers we didn't think they'd give it to everyone what
52. 7:18are we to do um so but I think that was on purpose but all of this is is all of these companies now going we've got AI
53. 7:26we've got AI oh we're going to be leaders too we've got a big model we've got a bigger model all that Panic is
54. 7:32because of chat gbt but the question was like how expensive is it really [ __ ] expensive now one of the cool things
55. 7:41about the fact that meta released all their large language models that like the the base called what what are called
56. 7:49the weights basically the The Secret Sauce the fact that they release those into the world the developer Community
57. 7:56The open- Source Community is pouncing all over these things things and they're now able to train for about $100 in
58. 8:03compute cost and get results almost as good as Google doing a foundational large language model that takes like 6
59. 8:11months and $10 million the developer Community is getting those things to similar levels of performance for like a
60. 8:17hundred bucks so it's like it's [ __ ] wild it is insane right now it really is just nuts so anyway um if if you're new
61. 8:27here uh put any questions or comments or thoughts or maybe even something you're trying to work on in the comments below
62. 8:34I'll do my best to answer them I'm going to flip cameras here and pause so I don't break but I'll show you these oh
63. 8:41we don't have that many people here so um go play with those Ur URLs over there um while you listen to me ramble if you
64. 8:50haven't played with chat GPT it's actually really important for all of us and I like I believe that deep deep in my
65. 8:58soul uh when will office when will Microsoft integrate GPT into office I I would assume as soon as they're ready like like Microsoft is
66. 9:09is they're [ __ ] acting like apple like like like I've hated Microsoft for decades like I was a Microsoft Basher
67. 9:17for [ __ ] decades we used to make fun of them in the early days of the worldwide web I had this group I started
68. 9:23called the worldwide web artist Consortium and we would meet at the Sony headquarters in New York City and and
69. 9:31Sony Sony gave us this it was this conference it was this boardroom but it sat like 200 people it was this massive
70. 9:38Square table and we'd all sit around this table it was like a hundred people would sit around the table and another
71. 9:44hundred were around them and we would have industry people come in and they would present like new web Technologies
72. 9:52and every time Microsoft came in we just [ __ ] annihilated them [ __ ] annihilated them they were just they
73. 9:59were just hated but but what they're doing with with this chat GPT and with with open AI is it's it's [ __ ]
74. 10:06brilliant um so probably as soon as they can um being Microsoft they're not going to they're not going to release
75. 10:13something that hasn't been really well thought out and I think how you actually integrate the co-pilot into Excel is
76. 10:21going to be different than how you do it in word and and you know is it always there is it like clippy that just
77. 10:27[ __ ] you know shows up annoys you like like they they seem to I I don't know what's going on with their product
78. 10:35development over there but their their product development seem seems to have gotten sort of a a step change better
79. 10:45than it's ever been like like their interface and their product has always been just over bloated horseshit right
80. 10:53we all have to use it cuz it all does what we needed to do but like it's just been bad what's what's beenen happening
81. 10:59lately is is they've been introducing these new tools like it feels like they've got some new product person
82. 11:05there or product team there that's that's dramatically improving things so as soon as it's ready I'm sure they they
83. 11:11they want to get it out before Google does and I think they will so I'm guessing fall you know September October
84. 11:24I don't know I don't have a [ __ ] clue but I want it I wanted code interpreter and we got code interpreter
85. 11:31yeah how much do it cost for all these big for all these big computers like like staggering amounts of money again
86. 11:39oh that oh that was the thing I was saying so Microsoft you know they went from 1 billion doll to1 billion and then Sam Alman has said to
87. 11:49get to AGI artificial general intelligence was that that's the goal for open AI is to get to a system that's
88. 11:56smarter than humans that can reason and you know basically be smarter than than the smartest human that's their goal and Sam Alman
89. 12:04said he thinks it's going to take them a hundred billion doar to get there and and that's going to be with you know all
90. 12:09of these tools out in the world and being used so it's it's going to be a lot it's a lot of money a lot of money how can I stay active on
91. 12:18teams I don't know show up to practices kick the ball do laps do push-ups Oh you mean Microsoft teams I don't know log in I don't know what
92. 12:31you're talking about I am looking forward to open AI in the XX what's the XX cross
93. 12:48pixs that is wait that is me that dude I don't know what you're saying there I don't what is XX is that some something that
94. 12:59an old guy wouldn't understand are you trolling an old guy by using youngster terms you son of a [ __ ]
95. 13:08I all right we need more likes to boost live guys support by tapping the screen oh thank you very much yeah feel free to
96. 13:15tap feel free to share the live anything we can do to let people know are here um I really do think it's important so so I
97. 13:22know I'm a little smarmy about this and I've been doing this for a while so I'm getting a little horse and you know numb buted
98. 13:31um and and a little Punchy um but but the the the real point of this channel is to demystify this AI stuff and get
99. 13:40people to try it get people excited about it and not excited about it like you have to love it like I don't give a
100. 13:47[ __ ] if you love it or hate it but I don't want you to hate it if you've never tried it like the whole point here
101. 13:53is learn this [ __ ] enough to at least have a Kevin McAllister moment which for young people there's this movie that us
102. 14:03old people saw when we were younger called home alone and Kevin McAllister did this once and when you use these
103. 14:12tools at some point it's going to do something that makes you do that and you owe it to yourself to use these tools
104. 14:21until you've had that moment and then at that point you can say [ __ ] it this is all [ __ ] right I think I think the
105. 14:29robots are going to kill us I'm just going to sit back I'm moving to the beach um or you'll go the other way and
106. 14:35you'll go oh holy [ __ ] [ __ ] like uh does this mean I should probably start learning this like right now
107. 14:43yeah because AI is not going to take your job but someone using AI will every job every knowledge worker job every creative job I I I can't
108. 14:57foresee a job that uses computers that is not going to be profoundly impacted by this [ __ ] and many jobs that don't
109. 15:05even use computers that are impacted by people using computers profound profound changes are coming from this so that's
110. 15:12what this channel is all about um what's nuts I'm nuts hey there hope you're having a great evening thank you very much
111. 15:20silhouette I appreciate that do you think oh Tik Tock just jumped me to the bottom oh tick tock oh here we go there's a danger of using
112. 15:31this do I think there is a danger of using this in the future danger of using what in the future AI oh like is there a
113. 15:39danger of using AI yeah it's so 3 days ago open AI made an announcement that they're dedicating 20% of their staff's effort that's a lot for
114. 15:53for a company that's doing as much as they are they're dedicating 20% of their staff effort to what they're calling super
115. 16:01alignment alignment in AI means are the goals of the AI aligned with human goals cuz if they're not and this [ __ ]
116. 16:18gets smarter than us and it and it doesn't have alignment and it's like you know I know those humans want me to not
117. 16:28shut down nuclear power plants and I know they want me to not do this but I think it might be in my best interest
118. 16:35for my survival as a system to do whatever it's going to do so we don't want that so so so openai if you want to
119. 16:45read about it you can go to um open.com blog and I think it's it's it's either the last thing they did or the
120. 16:52one before but it's this thing on super alignment you can go read about it it's it's fascinating but they're basically
121. 16:57going to use chat GPT to do all sorts of simulations and and do all these things to make sure that these tools are
122. 17:04aligned with our values as a human species so there is a danger because these things when they start executing
123. 17:11code which code interpreter executes code now it's got safety guard rails and things like that but again you've got
124. 17:18this whole open source community so there's a whole Community out there with these large language models without
125. 17:23safety guard rails on them right so that's happening in parallel with a company like open AI saying hey we're
126. 17:29going to dedicate 20% of our staff to the safety piece piece of this so um yes it's possible but I also think that
127. 17:39another reason for the channel why why I do these and why I'm so passionate about it is I firmly believe
128. 17:48that even small individual efforts on all of our parts to do something good with this technology anything we do to
129. 17:56bend it toward the light Bend end it toward the good right now is going to have a disproportionately positive
130. 18:03impact on the direction of this stuff in the future if we all just sit back and go ah I've got three years I don't
131. 18:10really need it's it's probably hype someone told me about blockchain and nfts and I I lost my grandma's money and
132. 18:16so this is probably just one of those anyone that's the more people that sit back and just do that and dismiss it and
133. 18:24and kind of live in fear of it um it ain't helping right but everyone that goes they have their Kevin McAllister moment they're
134. 18:34holy [ __ ] I can do that can it do my job here it oh [ __ ] what do I do now you learn you get curious you start
135. 18:44exploring oh well maybe I could do this maybe I could do that and then people have these breakthroughs where they're
136. 18:49like oh I know what I could do and I'm I'm seeing it consistantly people are getting their heads around this [ __ ] and
137. 18:54they're Reinventing themselves and they're like making these really cool positive positive trajectories and
138. 18:59they're getting really cool jobs so it it's super important uh am I excited about code interpreter [ __ ] yes now I got it today so so I was
139. 19:10actually speaking in front of a oh this was wild so today I did a an hour and a half I did a 90minut lunch and learn for
140. 19:19this New York City Data visualization company right it's a it's a good friend I used to work with at agency.com and he
141. 19:26invited me to speak to the to the group to his company and they do data visualization five minutes before that
142. 19:33call I logged into chat GPT and I said I wonder if I've got code interpreter and there it was so so I hopped on this call
143. 19:41I did my sort of talk for for 90 minutes and then the last five minutes I said let let me let me show you something and I showed them code
144. 19:49interpreter this is a company that does this [ __ ] for a living and they were like they were just like [ __ ] blank blank stairs like
145. 20:00because it was funny like when I started the call I had everyone go around and say what they were doing they were like
146. 20:04well yeah I think AI is okay but I'm taking a little wait and see attitude they all had attitude they were New York
147. 20:10you know New Yorkers I love New Yorkers they had a little attitude by the end of the call they were like uh uh I think we
148. 20:17got to deal with this so that was pretty fun but yes I'm very excited about code interpreter I just I saw I saw a Reddit
149. 20:24post um right before I came live that that basically said any anyone else feeling stupid today about code interpreter cuz it's like
150. 20:34it's like I don't know [ __ ] about data analysis so like I don't know how to get data sources once I get a data source I
151. 20:42don't know what to do with it I don't even know what questions to ask it so I'm feeling I'm feeling kind of
152. 20:49like I felt when I was trying to get stable diffusion running for the first time where I just felt like a [ __ ] [ __ ] for like
153. 20:56weeks and then I and and then he did the same thing when chat GPT came out I'm like I like I was doing parlor tricks
154. 21:02like having it write poems and songs and I'm like it probably does more right and I just felt stupid for like
155. 21:09months and then and so now I feel the same way with code interpreter but yes I'm very excited about it does chat GPT
156. 21:16read tarot cards it does and in fact let's go hold I'm going to show you a cool thing
157. 21:30we're going to go um chat GPT tarot card with images a prompt that generates tarot card reading so if
158. 21:46you uh let's see what site is this is this it yes chainbrain ai.com chain chain brain AI you can make money
159. 22:00with CHP come to chain brain AI you can make money with ch okay so this is just a couple of interesting prompts but one
160. 22:09of them is tarot card reading with embedded images so I'm I'm going to show you the prompt well I'm not going to show you
161. 22:15the to it there we're going to go to chat GPT we're going to play with it oh do I need no I don't I don't think I
162. 22:26do we'll do gp4 I don't think I need it connected to the internet to do this or maybe I do let's see okay so what this what this prompt
163. 22:38does is it says you're going to act as an expert tarot card reader generate a tarot card reading with tarot cards from
164. 22:44major and minor minor Arcana follow the following steps random number between 1 and 78 and then it's like insert random
165. 22:52number here insert the name of the tarot card that responds to that number and um do not use code block and then it it
166. 23:00points you to a a a Wiki media thing so what this prompt has in it is um all of the names of the files on Wiki media for the
167. 23:15images and then you basically just hit return and it pulls your tarot cards for you and and displays the images it's not
168. 23:29generating these it's literally what the prompt is doing is it's using markdown to point to a web image and then it's
169. 23:35just displaying the web image in your browser so it's not actually doing the stuff but it's a tarot card reading right it picks three
170. 23:44cards and what's cool about this is you know you can go back to that original prompt if you know anything about tarot card
171. 23:53reading you you could in instead of doing like a three card pick you you could do a nine card pick with two from
172. 23:59the bottom of the deck and you know all the sort of fancy [ __ ] you see on YouTube um so you could really bling the
173. 24:06[ __ ] out of this prompt and then yeah and then it pulls it pulls these images and if you don't like these images I bet
174. 24:13that there are other images somewhere else on the internet like there's there's is at Wikimedia right which are
175. 24:20kind of Open Source images but if you found another place on the net that had different kind of tarot cards you could just swap out this
176. 24:28URL with with the the sort of names of the individual cards down below isn't that wild like that that's the sort this
177. 24:36is the sort of [ __ ] if like if you want to if you want to like really start to get your head around prompt engineering and stuff like
178. 24:44this this is why you hang out on Reddit in the in the chat GPT forums because when when Geeks get their [ __ ] heads around what's possible
179. 24:56with this there's like like the the the depth of the rabbit holes that you can run down with chat GPT are [ __ ] insane all right I hope I'm
180. 25:17back why are you so horse Kyle I don't know maybe it's because I talk on the internet for 3 hours straight why can't you walk straight my
181. 25:27ass is numb you thought that was just a joke Liz does chat jpt read tarot cards you didn't think I'd have a [ __ ] answer for that one did
182. 25:42you it does I used to get on y'all's BBS which BBS back in the olden timey days on Echo all right we need more oh more
183. 25:58likes to boost the live guys support thank you for that uh boost by tapping the screen tap away people I encourage
184. 26:05the tapping I like the tapping the tapping is good no one's giving me a damn cowboy hat tonight what am I supposed to
185. 26:12do do you think this is all caps people do you think we will be able to download an AI software and use on our own GPU for
186. 26:21processing yes finesser we will be able to download on our own GPU okay so here's what's happening right now if you want to run
187. 26:31something the size of chat GPT you need some [ __ ] Hardware like George Hots the guy that from Kama the uh the the full
188. 26:44self-driving car competitor to Elon Musk he's a he's a weird but incredibly smart dude he's started a company called Tiny
189. 26:55Box where he's selling he he he's designing right now $115,000 computers that have six gpus in them so you can
190. 27:03run this [ __ ] natively so if you got 15 grand yes you can just [ __ ] fire that [ __ ] up um what's happening in the open
191. 27:13source Community is they're taking much smaller models so so uh GPT 3.5 is 175 billion parameters I don't know what the
192. 27:22[ __ ] that really means it means it's big it's a lot the llama models that were released into the wild the the the core
193. 27:31models that meta released there was a 7 billion a 13 billion and then a 35 billion or 30 I don't know like whatever
194. 27:3965 billion whatever they they're they're smaller what the op Source Community has done is they have trained those smaller
195. 27:47models by getting chat GPT to create training data and then and what's happening is chat gpts 4 is creating
196. 27:55this training data like lots of it they're training these small smaller models and these smaller models are
197. 28:01getting very smart very fast for like a 100 bucks in training what the developer Community is now doing is they're taking those models
198. 28:08that have good performance and they're optimizing them they're figuring out how much data they can pull out and still
199. 28:13have these things function at a high level so I saw so so there are now open- Source models you can install on a phone
200. 28:21there are open source models that run natively on an M1 Mac and an M2 apparently the M2 Macs are [ __ ]
201. 28:27destroying it um so so I think we are probably easily within a year of being able to do that relatively um seamlessly so yeah it's
202. 28:43coming U but if you want to run this big [ __ ] the big models you got to have some you got to have some iron as they
203. 28:53say chat GPT is 70% nerfed I don't uh disagree with that I think we we are not see well so like there's a bunch of
204. 29:02stuff you can't do um you can't do the image generation it doesn't have the 32,000 token limit there's there there's
205. 29:09a bunch of stuff that I have a feeling we're not seeing a lot of what this model's capable of I I think you're
206. 29:23right and I think they're slow rolling it on purpose I think they know that that they h i i I think they know that
207. 29:32they have to have it in the wild long enough to discover [ __ ] like they discovered with browse with Bing so
208. 29:39so if you don't know this three days ago they pulled GPT connected chat GPT connected to the internet they pulled
209. 29:46that functionality because what they learned probably from a lawsuit or a nasty call from a lawyer was that chat
210. 29:52GPT was pulling content from behind pay walls without paying for them Publishers apparently get a little cranky when that
211. 30:00happens um so yeah so I have a feeling they're trying to figure all this [ __ ] out before they really launch the the
212. 30:07powerful version of this and who the [ __ ] knows what chat gp5 is that what the core model of GPT 5 is it's going to be insane um you
213. 30:19switching your favorite for the first company to make a personal agent ubiquitous you switching your favorite for the first com oh
214. 30:31oh am I like will I switch to anyone that makes the first personal agent ubiquitous I'll switch to anyone that
215. 30:38makes a personal agent functional absolutely in fact this list of like the only reason this list I mean if you've been watching my
216. 30:49lives I've been doing these for what a month something like that this list hasn't changed I will change this list
217. 30:54the minute someone topples gp4 it'll go to the top of the list um the minute someone comes up with an autonomous
218. 31:02agent that doesn't suck that's that's actually a personal agent like that that Pi thing I'm I'm keeping my eye on them
219. 31:09I've got I've got my eye on anthropic and you know anthropics here in po um and and inflection I think it's
220. 31:17called Uh that's got this Pi personal assistant um the pi personal assistant I think it's called I'm keeping my eye on
221. 31:25them cuz that's the guy that founded Deep Mind and deep mind's the one that did Alpha fold that is [ __ ]
222. 31:32predicting um protein folding structures with remarkable accuracy in the hundreds of thousands like Wicked smart folks so so yeah if if
223. 31:45you're asking will I change my alliances in a heartbeat [ __ ] yes all this stuff is Chang like everything that's
224. 31:52state-of-the-art today is going to be laughably bad we're going to look back at it and go remember when we had to
225. 31:57type prompts in the chat GPT that's hilarious the kids will be like you used to have to type what are you talking about
226. 32:09Mama co-pilot is already in VSS code and it's pretty good interesting very cool laughing my ass off I caught myself I
227. 32:18caught up myself by commenting good why Nerf the system Xbox um AI versus educ oh versus Ed Educators or our universities done
228. 32:34um I think universities that pretend that this [ __ ] isn't transforming education are done I think any
229. 32:42Universities that are not like oh we got to change [ __ ] everything I think are done are in trouble yeah I
230. 32:51do any University that's still got a you know you can't cheat you know we're we're going to have anti-chat GPT
231. 33:01sniffers are done if they're if they're not if they can't look at this and recognize everything's fundamentally
232. 33:10different they're they're [ __ ] um what was it last week Harvard announced that their their sort of Flagship cs50 their
233. 33:19computer science their programming um like their entry-level programming class at Harvard is going to have a chat GPT teaching assistant
234. 33:27so you'll have 24/7 access to an AI teaching assistant for the computer programming course that will check your
235. 33:35work help you with documentation debug [ __ ] and then if the AI teaching assistant can't help you I suppose you
236. 33:45can go to office hours with the real person with the real Professor so that like that's Harvard leaning in the the
237. 33:52New York City School District initially banned chat GP and reversed that the city of Boston put out guidelines for gener generative AI
238. 34:02so you're starting to see there there was a there was actually a newspaper article in the New York Times probably in January or
239. 34:11February that interviewed a bunch of University provosts and you know curriculum people um that were actually
240. 34:19talking about we've got to take this seriously we've got to rethink education so there's actually you know there's
241. 34:24movement in the in the higher space to deal with this but there's also just like staggeringly idiotic close-mindedness to
242. 34:34it and you know that that's just universities are big institutions and big institutions hate change and
243. 34:43this [ __ ] is forcing change way faster than I think any of us are prepared for way faster I'm amazed at VBA coding through
244. 34:56open Ai and other programming languages wow I know it's amazing isn't it I know I know I know this what I'm this is
245. 35:05like have your Kevin McAllister moment right like use these tools until you have that moment where you're like here here here's the
246. 35:15thing chat GPT the the box that you type in looks like the Google box this ain't Google it's not a Google search
247. 35:28in fact it's kind of the opposite of a Google search where Google wants you to be as concise as possible just give us
248. 35:35the minimum amount of information you need in that box and then we'll figure all the other [ __ ] out chat GPT is kind
249. 35:42of the opposite where if you give it too little it just gives you crap back if you start interacting with it not like a
250. 35:49search engine but more like a collaborator like a like an intern like a really [ __ ] smart intern that that's clueless about
251. 35:58life that's chat GPT brilliant like Savant level brilliant knows all the [ __ ] and literally can't figure out how
252. 36:08to like take the next step so act like someone who knows what they're doing okay I want you to okay here you go oh no that's
253. 36:20completely wrong okay sorry about that want me to do it again it's like a simpleton chat GPT is like a simpleton that knows [ __ ]
254. 36:34everything really is and like the minute you the minute you sort of figure that [ __ ] out like something
255. 36:42shifts you're like holy [ __ ] I've got like a savant idiot partner I don't know it's I don't know it's good it's fun it's good it's
256. 36:57awesome by the way there's 100 people in here it's Friday night shouldn't you be out like drinking or dancing or clubbing
257. 37:05or something not sitting with like oldie mcart face in the in the in the in the AI in the in the in the smartass MC AI
258. 37:15mcon all right I'm trying to find where I was in the comments because Tick Tock is it's a conspiracy theory where Tik
259. 37:24Tock doesn't want me to answer all of your comments but I refuse to let the man keep me down all right there's the VBA
260. 37:38comment your broadcasts are fantastic thank you Firefly I appreciate that alignments is what humans have to do
261. 37:45that's to a great degree that's right right this is not it's one of my pet peeves is like people are like the
262. 37:50robots are stealing my [ __ ] the robots are doing this the robots wrote this the robots wrote that the robots lied no
263. 37:58human beings asked the robots to do something and then the human beings didn't check their work and the human beings are
264. 38:06idiots so so yes I agree you know humans need to get their [ __ ] together and get getting AI literate is probably the
265. 38:16single best thing any of us can do for the next year as as scary as it is and as technical as you think you're not you
266. 38:25don't have to be technical with this there's there's there's a bifurcation that happened on November 30th before
267. 38:33November 30th if you wanted to play with machine learning you needed the propeller hat and the math degree right
268. 38:41and you needed to be a coder and an engineer and you needed to live over in that side chat GPT marks an inflection point in the history of
269. 38:51humanity that we all anyone who can basically get onto a machine if you can get on to Tik Tock that's way more
270. 38:59[ __ ] complicated than getting onto chat GPT so you can get on to Tik Tok you can go here and we are like this is
271. 39:08the way I think about it I feel like I feel so incredibly lucky to live at a time in history where we get to play
272. 39:17with tools this powerful like we meaning everyday people not machine engineers and scientists all of us so wherever you are on the spectrum
273. 39:28of geeky to clueless you can do this prompt. chat will teach you how to do this chat. open.com is free bing.com is free po.com is
274. 39:41free go play with those things until it blows your mind and then get then dig in only people have access to the
275. 39:52unregulated version can speak about this I don't know what that was in references to um humans will have to change humans are
276. 40:00going to have to adapt absolutely Skynet okay somebody saw a movie Sky not the robots are going to kill us the robots
277. 40:08are going to kill us I'm sorry Dave I'm afraid I can't do that they might kill us what if they make make life dramatically
278. 40:22better that is a plausible future as well it's just that's a more boring movie so that's why we see 500 movies
279. 40:32where the robots kill us and like six where the robots are sweet nobody wants to see a movie where the we we hug and the robots Usher us
280. 40:43into a a a a great Renaissance of of creative expression the likes of which the world has never seen no one would want to see that
281. 40:53movie it's a possibility though if we choose to make it so in the mental health field I'd give it a Master Level PhD understanding
282. 41:05interesting I think there's something there's something really like like just kind of ironic and weird about the fact
283. 41:12that the large language model literally doesn't know what it's saying it's it it is it's got the it's it's got
284. 41:20the each word that it types it's doing it sort of with the functional equivalency of a c calcor right so if
285. 41:27you type in 4 plus 4 in a calculator and it spits out four that's what a large language model is doing word by word
286. 41:34it's just spitting out a probability this word then this word then this word then this word it's not aware of anything it's writing
287. 41:42but it's so good because it's been trained on what it's been trained on and they've got it fine-tuned how they've got it
288. 41:49fine-tuned it's it behaves remarkably well with empathy and with nuance and with engaging so so yeah I think
289. 42:01companionship and um therapy a and just mental health in general and and caregiving in general I think that these
290. 42:11large language models are going to be [ __ ] brilliant I think they're going to make doctors better I think they're
291. 42:17going to make caregivers better I think they're going to augment the health care industry and yet they're completely
292. 42:25clueless but there was a study I don't know a month or so ago that a bunch of patients interacted with chat GPT
293. 42:32doctors you know via text and real doctors via text guess which one they liked more it wasn't the real
294. 42:41doctors more competent more empathetic that was the feedback the chat GPT doctors more competent more empathetic it's like holy
295. 42:53[ __ ] maybe these tools will do all the shitty ass Insurance paperwork that makes doctors so cranky and makes them not spend time
296. 43:05with patients so maybe it will free them up to spend more time with patients and maybe it will actually train them to be
297. 43:11better doctors like better bedside matter better empathy I think that's coming I got my floating screen back I
298. 43:21know I did they gave it back to me I was a good boy oh [ __ ] I have to see if if I can enable code interpreter yeah go go
299. 43:28go if you've got chat GPT plus go to your settings and go to Beta beta apps or beta beta features and see if code
300. 43:37interpreter is there and if it's there click on it turn it on go play I'll I'll show you how to do that here in a second
301. 43:44if you want um expand what did you do expand what did you do what did you do what did you do bad dog bad dog I don't
302. 43:56know what I did but I was very very very bad it's amazing when you talk to people about AI there's a real division between
303. 44:05people on AI there is that's why this channel exists that's why this channel exists thanks Ron um everyone that's pissed off about it
304. 44:17that I know and everyone that dismisses it and everyone that demonizes it hasn't [ __ ] used it it's why I start out and mention 30
305. 44:31times a [ __ ] live go here and play with these tools if you haven't played with chat gbt go here and play with that
306. 44:40until you understand that this is not a Google Search and this is not a toy and it's not hype it is something
307. 44:54different it is a reflection of humanity that we get to tap into and it's a Humanity amplifier it can take this [ __ ] can
308. 45:05take your little Whisper of an idea that you're like you know maybe someday I'll fill in the blank write a novel write a
309. 45:15play write a recipe start a garden start a community group start a church whatever the [ __ ] you want to do
310. 45:23that little whisper that you've always kind of avoided cuz you're like I don't really know how to do it I don't know where to start go
311. 45:32there and just start [ __ ] with it if you haven't done it have it write a poem have it write a song have it write a project plan have
312. 45:43it write a marketing brief have it write a blog and then and then at that point you'll be like huh that sucks less than I thought it
313. 45:52would but surely it can't do what I do and then it will some version of what you do and then you should be freaked
314. 46:02out appropriately you should be absolutely [ __ ] terrified and you should be inspired to go holy [ __ ] I got to deal with this I got to
315. 46:13figure this [ __ ] out then We're Off to the Races so yes I agree so and and you know what I would say is for those people
316. 46:24that are violently opposed to this and angry about it like have empathy for them because it's [ __ ] scary and and if you
317. 46:36know especially for people that you know are on in years and maybe have 20 30 40 years into a profession and all of a sudden [ __ ]
318. 46:50like me are going you might want to reconsider your career choice well I'm [ __ ] 30 years into it jerk so you
319. 47:01know you know have empathy and then I think it's I you know I think the thing to do is is do everything you
320. 47:08can to make what you know about this stuff accessible to them in a way that it you know they can sort of see themselves in in in it right do
321. 47:21something for them say Hey you know weren't you trying to to come up with some book ideas well yeah but I didn't
322. 47:30really get it started hey you want to you want to try something what what was the what was the idea about type that
323. 47:36into chat GPT did do you have an outline well I don't really know how to do an outline let's let's see if we can get this thing to make an
324. 47:42outline you know give me three different story Frameworks that I could you know write my novel around okay here they
325. 47:51come oh wow and and and what's the character's name name Bob okay let's put Bob in there and what's what's Bob do is
326. 47:59it mechanic okay we'll put that in there write me the outline of Bob who finds his true love you know at the ball
327. 48:07game and whatever the plot points are out comes the outline then give that to them that's pretty cool that's that's
328. 48:17not as scary as I thought you know try to get people there open AI just posted positions for their alignment Team all north of
329. 48:28$400,000 there you go wow Jesse really holy [ __ ] $400,000 starting starting at $400,000 now if you're going to go work
330. 48:40at open AI for their alignment team I assume you got to have some [ __ ] Tech chops I assume you got to have your [ __ ]
331. 48:46going on with this here uh this here uh code interpreting but I don't know it could just be Charming good haircuts that get you in
332. 49:00there I guess I'm out [ __ ] I had this idea which one laughing my ass off at the mental image of Kyle
333. 49:08deep in the YouTube rabbit hole of tarot card dealing exactly can you share this can I share what I'm sharing everything I don't know
334. 49:17what uh Joe famous good name by the way Joe famous that's solid I like it and then right below Joe famous is Johnny
335. 49:24bullets this this is awesome tonight except for the troll that told me I was going to be permanently banned Jesus
336. 49:31can't believe I fell for that what next a Nigerian prince oh that's the other thing the Nigerian prince letters are going to get
337. 49:38much better written you know that right okay um so I don't know what you were talking about Joe famous I I can share
338. 49:45whatever if I know what it is um where do you go to make it to video for free to make what to video to to like do text
339. 49:55to video video um you can go to Runway ml um are those free I think they are free I don't remember anyway go to
340. 50:07Runway ml they have an app and then they have a Discord server where you can generate you can either do video to video or text to
341. 50:17video I I think that's what you're asking Johnny bullets but I'm not sure what are the cool what are the cool jobs out there for AI newbies
342. 50:32um I think every single company within within probably before the end of the year but you know let's let's call it
343. 50:40within a year let's get let's give people another year we're in July let's say by this time next year every single
344. 50:49company is going to need a significant amount of AI literacy in in in their organization um and they may already
345. 51:00have it but not know it um go go check out there's an article from a guy named Ethan mullock called detecting the
346. 51:06secret cyborgs where he's talking about there's a class of AI literate people in companies right now that are using AI
347. 51:14for personal productivity like dramatic improvements in personal productivity but they're not telling their companies
348. 51:21about it they're the secret cyborgs and the companies are like well these tools aren't ready for prime time and we don't
349. 51:26have a privacy policy and we're not really sure what we're going to do with AI and so the organizations are are kind
350. 51:32of like trying to figure out if and when they're going to start doing this meanwhile they've got a whole class of
351. 51:38employees that are AI literate and getting increasingly so um and and he's saying basically bridge the gap
352. 51:47so so what I would say is rather than what are the cool AI jobs out there you know there's there's going to be prompt
353. 51:54engineer jobs and there's going to be just all sorts of thing I would say go back to yourself what are you good at
354. 52:00what are you passionate about get AI literate and start talking about it start talking about it on LinkedIn start
355. 52:07talking about it on Facebook start talking about it at [ __ ] dinner parties start annoying people like
356. 52:13people can't stand talking to me anymore I'm convinced of that cuz all I [ __ ] talk about is this stuff why cuz it's
357. 52:20going to change the world and I want to be a part of it and I want to bring people along it's a [ __ ] adventure to me me um
358. 52:28in getting AI literate and in communicating it with the world you will become known within your circles as one of those people that when
359. 52:41they finally when the light bulb finally goes off for them and they're like I got to learn about this AI oh you know what
360. 52:49um what's your name Saku Sal Saku um Sally knows about that Sally's been talking about that for like two months it's all it's all they
361. 53:03talk about hey Sally Could you teach me this like be that person so so so I don't know I mean start paying
362. 53:13attention to job boards um within the AI Salon the the salon group that I created um a number of people within there came
363. 53:24into the group not knowing anything about AI other than having a um curiosity about it started learning
364. 53:33about it started posting about it put it on their LinkedIn profile started posting articles and posting results
365. 53:40here's something I did in mid Journey here's something I did in chat GPT are getting jobs are getting cool
366. 53:46jobs are getting AI Centric jobs so I would say just start learning start putting yourself out there I'm already
367. 53:54running models on my RT TX 490 that's super cool thanks Ron Burgundy that's okay solid solid troll compliment sort of Burn yeah I like
368. 54:08it don't Google TIN Tiny package Tiny Box George hot's company is called Tiny Box not tiny package you should definitely do not Google tiny
369. 54:21package don't do I know you're doing it right now you're all doing it right now you're Googling tiny package don't do it Jesse tried to warn
370. 54:31you GPT for video free no BS I go try Runway ml but also just go to Future
371. 54:54pedia.com imagine any are are free free maybe maybe you can get some amount of credits each day what are the names of
372. 55:02those open source models oh the the the core models that meta released are called llama and then one of the first ones
373. 55:11that was sort of trained up was called alpaca and now there's one called vicuna and then there's one called MPT 7B and then there was MPT
374. 55:227B which stands for 7 billion an MPT 7B Storyteller or story something that had a 65,000 token context window and
375. 55:36like what I would do is I would I would just go to Twitter and to Reddit and type in op Source llm both on Twitter and Reddit and just
376. 55:49run down those rabbit holes I can't keep up with the open- source models anymore the the base mod model that meta
377. 55:55released are called llama like the animal um and then all these open source projects are stacking on top of it and
378. 56:04there's just there's hundreds of them now um and all sorts of different flavors of them it's just it's the open
379. 56:11source community and it's it's the open source Community where everyone has basically you know dived in dove in dove
380. 56:19in head first cuz this AI stuff's so exciting um so just go to Twitter open source llm and redddit open source llm
381. 56:28and you'll you'll find you'll find projects there that are people are excited about they probably already have
382. 56:34five if you COI if I can explain to me what chat GPT actually is I'd really appreciate it um all right how about I
383. 56:44teach you how it works you want to know how chat GPT works well so let me explain what it is okay now let me h
384. 56:57I'll show you how it works in a second [Laughter] [Music] but so you know how Google
385. 57:15is they indexed all of the internet they went out and they made hyperlinks to websites and they said this hyperlink
386. 57:25points to a website that talks about this [ __ ] and then when you go say I want a website about that [ __ ] Google
387. 57:32has somehow figured out how to search all of the sites on the internet and give you links that point you to those
388. 57:40sites chat GPT has learned not just the link and what the site's about but it basically goes behind it and it's learned all the
389. 57:50[ __ ] at the site so when you interact with it on its surface it feels like kind of a Google search box it's just a simple here hang
390. 58:01on let me show you it's just a simple blank window with a with a simple it says send a message right but the message I can send instead
391. 58:16of getting back a list of links to other sites I can put in here explain the main concepts of quantum mechanics as well let me just do
392. 58:33that the main concepts of quantum mechanics and so you see how fast it's writing so it it's writing now it's
393. 58:41explaining to me what quantum mechanics are it's not taking me to a site that's been SEO optimized it it it knows all
394. 58:49the [ __ ] it knows all the languages it knows all the concepts it knows all the [ __ ] so it's explaining to me
395. 58:55wave particle duality and superposition and Quant quantization and uncertainty principle and wave function and
396. 59:02probability and entanglement you know all the cool stuff with quantum mechanics but now I can say um explain it like I am
397. 59:16five sure imagine you have a toy car that can also act like a ball sometimes when you play with it the car moves like
398. 59:24a car and the other times it rolls and bounces like a ball so so in 10 seconds that was it it shifted from explaining
399. 59:32quantum mechanics to explaining it to a 5-year-old now I can say um make it an m m WP first one yo listen up
400. 59:47kid I'll break it down real quick quantum mechanics it's a mindbending trick picture a toy car it's got dual
401. 59:52identities acting like a car bouncing ball with intensities so it's rhyming and talking sort of like Eminem
402. 1:00:01would talk to a 5-year-old would wrap to a 5-year-old so now I can say um make a table that compares the
403. 1:00:15Einstein um relativity world and the quantum world and now it will hang there you
404. 1:00:31go it makes a table comparing two fairly deep subjects Einstein relativity relativity macroscopic Quantum World microscopic subatomic right
405. 1:00:46um write a python application application that um
406. 1:01:00takes some inputs and simulates both physics models sure here's here's a python script here's a python application that
407. 1:01:16allows you to input values and Sim simulate both Einstein relativity world and the quantum world the example
408. 1:01:23focuses on basic calcul ation and simulations rather than comprehensive models you can modify and expand upon it and then there's your code
409. 1:01:32so what I just did is what I consider the parlor tricks of chat GPT but if you think about what I just did and what it just did it is
410. 1:01:44staggeringly remarkable right this is not a Google search at all this is us being able to tap into to effectively the collective knowledge
411. 1:01:58that's been put on the internet for the past five decades the public internet anyway and do it instantly and do it
412. 1:02:06exactly as we want it and and get like be able to bend this tool to our will so so what it is is a tool that allows you
413. 1:02:15to kind of do anything you can imagine when it comes to knowledge you could learn something you don't have to learn
414. 1:02:22programming it will program for you it will teach you how to use its programming that it creates to actually
415. 1:02:29do something you want you know if you want to you want to be an app developer but you've never learned how to program you can actually do
416. 1:02:38that going to take you some work but it's kind of here to teach you to do that so the let me let me flip the
417. 1:02:53screens the thing that makes it more remarkable is how it does what it does I don't had it partner you know what we're talking about here oh
418. 1:03:05hey you know I kind of like that their internet you ever been on that internet they got this new thing it's called AI um
419. 1:03:21cheers the way it actually works is when you type in your prompt some [ __ ] Geniuses out there Jeffrey Hinton and the people that made
420. 1:03:32the Transformer there's a bunch there's a bunch of geniuses out there what they figured out they they figured out something that
421. 1:03:40you've already played with so the AI already exists in in almost all the software we have you know when you're in
422. 1:03:45Google sometimes you're writing an email and you write a sentence and Google will put those gray words at the end or
423. 1:03:53you're in Microsoft Word or where it completes your sentence that's essentially what chat GPT is but it's like that on
424. 1:04:01steroids so when you write your prompt and you submit it what it does is it turns that into a mathematical probability
425. 1:04:10engine that says okay based on all of what this person just wrote what is the next most probable word that is the ne
426. 1:04:18that's the right word and it and it puts one word there and then it takes all of that plus one word and it says well
427. 1:04:25what's the next most probable word and what's the next most probable word and it does that over and over and over so
428. 1:04:30when you saw it just spit out here's here's quantum mechanics explained here's the poem here's the code it's
429. 1:04:37doing that one token at a time they call it a token it's it's part of a word it doesn't matter but just think one word
430. 1:04:43at a time over and over and over again just a probability engine so so conceptually it's quite simple what it's
431. 1:04:50doing but realistically because it's been trained on [ __ ] everything and because it's so fast it's it's like this magic access to
432. 1:05:02the knowledge of the of humanity you know some large percentage of it so I hope that explained it the raw gp4 model was aoral but fast
433. 1:05:14and vastly smarter than it is today yep and it was well it was it was unremarkable it got remarkable as they
434. 1:05:20fine-tuned it and then they had to put guard rails on it to to you know keep it from doing some bad [ __ ] so
435. 1:05:28like I you know I don't listen as as hobbled as gp4 might be if you [ __ ] use it on a regular basis it's still
436. 1:05:37magic it's still a remarkably powerful tool like in the hands of just Everyday People this is the most powerful tool
437. 1:05:46ever put in the hands of normal people it's you know I I I say it on here a lot but I think it's Carl Sean was it Carl Sean
438. 1:05:54I think it's Carl San wait hold who
439. 1:06:12said Arthur C Clark any sufficiently advanced technology is indistinguishable from Magic that's what this is it's an advanced technology it
440. 1:06:25couldn't happen without these supercomputers and without the crazy ass math that some people figured out but what it does is
441. 1:06:35Magic you ask it to write a rap or a country song or your term paper or your email or your resignation letter from
442. 1:06:44your shitty ass company that you want to leave but you don't want to burn a bridge and it [ __ ] does it like
443. 1:06:50that and it does it good it doesn't probably better than you would it gets you to 80% in 30 seconds and then you
444. 1:06:57can take it for the for the rest of the 20% do you think we would be able to share GPU power to run GPT like people M
445. 1:07:09Bitcoin well that's that's exactly what Bing is and that's exactly what chat GPT is so all the free tools here chat GPT and Bing are
446. 1:07:21that because every queer that you do it's like it's burning you know multiple gpus to be able to answer your
447. 1:07:32dumb little question right my dumb little question like we we're doing that right now now are you saying that W will there be
448. 1:07:44Services where I can take a powerful open- Source large language model and and have shared GPU Bank so I don't have to send my data to
449. 1:07:56the man and to Microsoft and Sam mman yeah probably sure that'll come you know unless the powers that be
450. 1:08:04are like yeah we're going to we're we're going to cuz I think there was some legislation somewhere where they're
451. 1:08:09trying to lock down the sale of gpus we're not going to sell gpus to the undesirables which is basically people
452. 1:08:16we can't make money off of so we'll we'll see what happens with all the legislation and regulation stuff um how
453. 1:08:24do they know that chat GPT doesn't pull content wait how do they know that chat GPT doesn't pull content from behind pay
454. 1:08:31walls they know that it does that's why they puled that's why they removed it so so pay walls work there there's kind of
455. 1:08:38two flavors of pay walls so when you go to like the New York Times or you go to some site and it says sorry you've got to subscribe to
456. 1:08:47this one way you do that is kind of the cheap way and if someone is Technic Tech and they know how to inspect the code of
457. 1:08:55a website they can just inspect the code and there's all the content so the pay wall is just this sort of [ __ ] thing
458. 1:09:03there's another way you do pay walls where it's actually behind a pay wall and you can't just inspect the code and see it I
459. 1:09:11think browse with Bing if you had a shitty pay wall that just you could inspect the code and get the content I
460. 1:09:18think browse with Bing was just pulling that content and saying here you go but if you do if it's a big company
461. 1:09:26that had a shitty pay wall that big company can still get pissed off and say we're going to send all our lawyers
462. 1:09:31after you and destroy your company so that's that's my guess we will laugh that computers had keyboards yeah I
463. 1:09:38think so I mean I think if if you look at the technology in in the Vision Pro from Apple what's coming and and yeah although you
464. 1:09:49know I don't know that typing goes away because there's a lot of times when I'm in a situation and I don't want to um I don't want to
465. 1:10:01talk and I mean you know 10 10 years 8 8 10 years you can probably just think thoughts and that's your interface so
466. 1:10:11you don't have to talk you can just think it so that's probably coming don't get me wrong is the background on your laptop the Sam
467. 1:10:20Alman I orb it's actually not a bad that's not a bad thought this is an image um from from stable diffusion that I created I have an art
468. 1:10:35project called Kyle Shannon dreams and this was one of the images that when I created it I I had a visc visceral
469. 1:10:47Rea so it's like there's my you can you can see my chin my little double chin my lips my nose and then that thing on the eye
470. 1:10:57um this was one that that when it was created like when it showed up I I like had this emotional response this is the
471. 1:11:05kind of art I would create if I had the skills to create it from scratch and I don't but I don't and so when I saw this
472. 1:11:13I'm like that's that's that's [ __ ] magic like this is what I would create um so that's why I use it but yeah if
473. 1:11:20you don't know what he's referring to hang on let me flip back here so Sam Alman has this company called worldcoin and they have this
474. 1:11:32orb that you stick over your eye and it scans your iris and it verifies that you're human and if you're willing to do that
475. 1:11:43Sam is going to give you some cryptocurrency called worldcoin and it sounds creepier than [ __ ] but a lot of this stuff does and
476. 1:11:52what's going to happen what what the purpose of it is is imagine a Twitter or a threads you know in today's
477. 1:12:00parlons where every person on there is guaranteed to be a human and you know they're human and and I can say oh that's a video from the
478. 1:12:11verified you know Kyle and the verification isn't $8 I pay to Elon Musk but it's actually that's the verified
479. 1:12:20Kyle right and then you could imagine built on top of that could be here's a video that Kyle produced and and it's
480. 1:12:28blockchain verified that it hasn't been modified and here's an article that he wrote and here's a this and here's a
481. 1:12:35that and here's all the people that commented on it and those were the actual comments and nothing's been
482. 1:12:41manipulated so so the the orb thing is is a very real thing trying to solve what's going to be a very very real
483. 1:12:48problem that I think we're going to see in the 2024 election here in the states where uh some shit's going down where if you thought
484. 1:12:58misinformation and wacky High Jinks were big in 2016 and 2020 you ain't [ __ ] seen nothing 2024 is going to be insane with Miss
485. 1:13:10information um the era of human intelligence is already over I think we'll get it back is Google search dead
486. 1:13:16Google search is [ __ ] struggling I think they're gonna I think I think they're going to have to reinvent
487. 1:13:24what it is I do I don't think it's dead and I think it's I I think you know there's there's there's an there's enough of an
488. 1:13:35established infrastructure of search and SEO and websites that they're going to stay for a while but I think what
489. 1:13:44happens is the importance of destination website starts to plummet like right now Google's whole model is you search for
490. 1:13:52[ __ ] they give you you a link they want you to click on 30 of those links and go to a website what chat GPT is doing is
491. 1:14:00you ask for [ __ ] and it's there you don't have to go anywhere that's fundamentally a different thing it's just a completely
492. 1:14:09different Paradigm and so I think it makes the importance of a website less important for end users cuz end users
493. 1:14:18could just say hey my knees hurt oh um one of the things that might help is minimalist running shoes oh what are
494. 1:14:26minimalist running shoes here's what they are why do they make your knees hurt less here's the science oh that's
495. 1:14:32kind of cool who are the companies that do that here's the list of companies that do that oh can I what are the shoes
496. 1:14:39look like here's what they look like oh can I buy that one yeah here it is here's the link buy it click it's on its
497. 1:14:44way I never left chat GPT or whatever the tool is at the time it might be my agent that goes out and does that for me
498. 1:14:52there's no need for me to go to a website so if you play around with Bing what you'll see is they're starting to
499. 1:14:59incorporate you know citations and links and things into the answer so it's probably still going to have
500. 1:15:09like SEO but it's going to be more like llmo large language model thank you thank you thank you large language model
501. 1:15:17optimization how do you get your company to show up in that conversation about minimalist running I don't know someone's going to
502. 1:15:28have to figure that [ __ ] out Google wants you to still go through their portal to do all this but right now Bart
503. 1:15:35isn't as good as chat GPT and Microsoft is subsidizing chat GPT and Microsoft doesn't give a [ __ ] if they make money
504. 1:15:43on Advertising right now so it's in trouble you know what did they call they call it a red alert or something like that
505. 1:15:53like four weeks after chat gbt came out Google was like H I know we invented this [ __ ] but uh we're behind we're gonna have to do
506. 1:16:03something here anyway listen it's 11:45 here in Denver so if anyone's on the east coast God damn you need to go to
507. 1:16:09bed how do I feel about uh solidarity AI I don't know it I'll I'll look it up though for the guy who wanted to make
508. 1:16:17the video with chat GPT check out the real Melissa zando on Tik Tok okay cool awesome I love that you're helping one
509. 1:16:27another hard to hear with the music not sure what music we we got a bit of Jack Nicholson Vive here's Johnny yeah my hair's my hair's yeah
510. 1:16:43that's that's yeah yeah we got some jack in there today yeah yeah yeah now we got some jib carry all right Tom Arnold Vibe combo nice
511. 1:16:53Jack see these are the kind of trolls that I like what's the difference between Ai and a quantum computer um AI
512. 1:17:02needs computers to run a quantum computer is a quantum computer that can process ridiculous amounts of data
513. 1:17:10um quantum computers aren't quite ready for prime time yet but when they are uh you're going to be able to process a lot
514. 1:17:16more [ __ ] a lot more quickly so I would imagine that the ability to train these models goes up dramatically and they get
515. 1:17:25powerful to the point that they probably need to be licensed like you license a nuclear power plant because they'll be
516. 1:17:31that powerful that's what Sam alman's pushing for he's Sam alman's pushing for three tiers of Regulation the middle tier is kind of
517. 1:17:42where we are today gp4 Bard llama you know the these foundational large language models that are very very
518. 1:17:50powerful but they're kind of stupid he's saying those the companies that can produce the the base large language
519. 1:17:58models they should be regulated he's saying any of those companies that reach AGI or you know higher level cap
520. 1:18:06capability that should be licensed like a nuclear power plant and then what he's saying below that none of that should be
521. 1:18:12regulated all the small companies using these tools should be able to [ __ ] just go nuts and do what they do um so
522. 1:18:20it's going to be interesting to see how it plays out anyway everybody uh this was a good long night I'm a little horse my butt's numb
523. 1:18:29you know it's a good sign it's time to go can't feel my legs but uh it's been a pleasure happy Friday night or for happy Saturday
524. 1:18:37morning for some of you if you're on the East Coast peace out everybody I will probably see you sometime tomorrow I'll
525. 1:18:43be around you know I'll go fix the hair and come back peace